Wednesday, August 13, 2014

EVCCON 2014 - Day 2

Today is the second day of presentations at the A.C. Brase Arena.  We look forward to tomorrow at the airport for drag racing and autocross and dyno testing, then Saturday for 3 morning sessions and the car show in the park.

First Session:  EV Technologies And Racing Applications - Adam Clark and John Metric

Adam and John kicked off by showing their drag racing plasma video I posted yesterday.  Adam and John are leaders in the field of electric drag racing.  They met and decided to build a drag racing Miata and they've set a couple of world records and always beat the gas guys!  Their car is called Assault and Battery, best 1/4 mile time is 8.9 seconds at 149 mph, the fastest car with doors in the world.  Don Garlits is a legend in top fuel drag racing and has come over to the electric side, which is bringing a lot of attention, trying to be the first to hit 200 mph in an electric dragster after being the first to go 200 mph in a gasoline dragster.  John Waylan's car is called White Zombie.  His youtube videos are legendary.

Electric motorcycles are getting fast, 0-60 mph in less than 1 second, 0-200 mph in 6.9 seconds!  The fastest motorcycle at Pikes Peak in 2014 was electric.  The Nissan Deltawing car raced at Le Mans this year, along with the Drayson racing car.  Formula E is a new open-wheel racing series sanctioned by the FIA, the world motorsports leader, and I will be attending the second Formula E race of the season in Kuala Lumpur.

As with conventional racing, a lot of the innovations trickle down to smaller racing series like SCCA and into road cars, and we expect that technology from electric racing will make it to electric road cars too.  Advances in shedding weight by using lighter and stronger materials, advances in battery capacity and power delivery, supercapacitors, controllers, motors, and software for control and data collection will make our road cars better in the future.

Since motors are so key to racing, Adam went into great detail.  Series Wound DC can go very fast for a short period of time, so they're good for dragsters.  Heat and brush wear limit their more general use.  Brushless DC are more suitable for smaller vehicles like go-karts.  AC Induction motors are good for larger vehicles by choosing the desired torque curve.  They run higher voltages which improves efficiency to 95% but the controllers are more complicated and expensive and can't yet offer the power needed for drag racing.  Permanent Magnet AC motors are a little more efficient and a little smaller and lighter than induction AC motors but will demagnetize if they get too hot, so beware.

A study was done comparing AC induction and Permanent Magnet styles.  In the city test, the PM motor was 27% more efficient, but in highway use they were equivalent.

Popular DC racing controllers include the Zilla 2K with 2000 amps at 380V, the Shiva is 3000 A at 425 volts.   AC controllers include the Scott Drive and several Curtis units.  To get more power, you have to increase volts or amps, or cool the units better.

Where do the business opportunities lie around electric vehicles?  Material Science is a key job for the future, trades such as welding and machining, software development, carbon fiber, aerodynamics, energy management devices like DC-DC converters.  Lonestar offers their own DC-DC converter and AC motor controller with 1400A continuous and 2800A burst, up to 1200V in, up to 1.5 megwatt output!

Next Session: AC Drive Systems by Brian Seymour of HPEVS

Brian's father started an electric motor business by rewiring broken motors for friends and family, and the business has taken off.  They got into the golf cart business and pioneered the use of AC motors.  Brian's generation moved into motors for electric cars, delivery trucks, boats and other vehicles.

Brian converted a waterski boat from a gasoline engine to a newly developed stainless steel motor called the AC35x2, containing two motors on the same shaft and oil cooled.  They also supply mounting plates and couplers to Mercruiser drives.  It has dual Curtis AC controllers and 144 of CALB's new CAM72 batteries.

HPEVS offers a wide range of AC motors, either single or dual units on a single shaft.  Some of the motors are air cooled, but most are water cooled now, for the future they're moving to oil cooling.   They're using Dielectric transformer oil, which is really good at getting into contact with every part in the motor and carrying away the heat.  Water cooled motors are just jackets around the hot parts which can't carry away all of the heat.  It also acts as a lubricant for bearings and seals, and allows for higher operating temperatures.  The impact of using oil rather than water or glycol has an impact on the pumps, hoses and radiators used in the cooling loop - they need to be selected for oil and expected temperature.  Mercruiser hoses are readily available, and oil pumps can be found at Summit Racing.  They suggest using the existing automatic transmission radiator.  They've also added an oil filter for safety.

Brian then showed a video of a ride in his boat, nice and smooth and quiet!  He then walked us through the performance charts.  The big takeaway is that his boat was consuming just over 10 X the watt-hours/mile as we would expect in an electric car of the same weight.  This is why you should plan on a very large battery pack in your boat, or take short trips.  Note that getting up on the plane is key to minimize hydrodynamic friction, but there is a big power penalty trying to go past.  At 42 mph, the motor was pulling 1,000 amps!  Changing the prop to hit your target speed at the sweet spot of the motor will minimize your power consumption at your desired top speed.  The weight difference is a net gain of about 100 pounds, so it's almost equivalent of stock.

Next Session: High Performance from Stock Parts, Part 3

A followup session to the first this morning from John Metric, President of the National Electric Drag Racing Association.  John crashed his car last week doing an interview with MotorWeek on PBS, but was calmer than now standing in front of our friendly group.

This is an update from last year's talk.  He uses twinned Warp 9 motors like I have in my little 914, but he's pumping an insane 2000 amps of current through each, using Lithium Cobalt pouch cell batteries.  The batteries he uses have doubled in output current in 4 years.  He hasn't used a BMS for 1.5 years but is still interested in monitoring the each cell, since he only has 30 minutes between runs at the drag strip.  He's built a device that plugs into his pack and gives a quick view of the health of each cell, highlighting the high and low cell in the pack.

He showed several graphs overlaying battery pack performance and controller capability, showing where they match up nicely or are over or under capacity.  He then showed some simulated dynamometer graphs with torque, rpm and horsepower curves.  Air is a real performance problem, with enormous increases in horsepower needed to overcome drag.

What John learned about acceleration of a drag car, is the weight of the car and the road surface coefficient of friction is all that matters.  Torque and horsepower don't even come into the equations.  Another big factor is twisting the chassis due to the spinning motor, he was raising only one tire off the ground, so you have to pre-load the chassis so the car is flat as it's going down the track.

If you want to be entertained, find some youtube videos of electric drag racing!

Next Session: and Electric Boats by Anne Kloppenborg

Anne is the face of EVTV in Amsterdam.  He builds cars, converts boats and is a major EV component supplier for Europe.  He sends Jack a video update nearly every week, so you can keep up with his efforts by watching EVTV.  They've grown so much they're moving to a much larger space about a kilometer away in October.

His first build was a Glastron ski boat, and then moved to boats with less weight and more efficient hull designs.  They then did a Nedcraft Silverback, a beautiful boat.  Then they did the Ray Wright Delta with two drive train revisions.  Now they're working on a cabin cruiser.

He showed the design and construction of the New Delta, which is the basis for the boat that Anne built for Jack.  It's somewhere on the road between Chicago and Cape Girardeau right now.

They're working on a Jeep conversion in the shop right now, and have built a set of adapter plates that and reduction drives that will allow them to mate any supported motor to almost every boat.

One big issue they've run into is the need to pass very strict electrical interference certification testing, which happily they did for their DC motor system, and recently submitted the AC solution, working together with competing companies in the area.  Unfortunately they overheated a resistor but it was an easy fix.

Anne has a goal of crossing the English Channel with an electric boat.  The Guinness Book Of World Records has, at the request of his friend Al Gaida, created a new category of Fastest Crossing Of The English Channel By An Electric Boat.  They will make the attempt early next year.

We are now transitioning to a general boat panel session.

They discussed the cost of ownership vs. gasoline, electrical noise in the wiring, the motor required for a large Amsterdam tour boat, reduction drive belt power handling, converting a personal-sized watercraft and details of the channel crossing.

Next Session: JLD 505 with Paulo Almeida and Celso Menaia

Paulo and Celso are instructors at an engineering institute in Portugal.  They are circuit and circuit board designers and came up with one of the latest versions of the GEVCU boards.  They are also the team leaders on the flash-conversion of the Smart car this week.

In their spare time, they've developed what they're calling the JLD 505.  This is their take on a Chinese device called the JLD 404, which is very popular in the EV world as it not-very-accurately measures voltage, current, amp-hours and watt-hours and controls two relays which are triggered from thresholds you can set.  It has a large blue LED display with buttons to scroll through the various values and is attractive to install in your dashboard.  It is key to determining the state of charge of your pack which keeps it healthy and gives you as the driver the information you need to avoid running out of juice on the side of the road.

Paulo and Celso are introducing to the world the first prototype version of the JLD 505.    It offers a better display and data output.  Their design protects the internal components with a fully isolated power supply, USB port, Bluetooth and CAN bus, two digital inputs, two digital outputs, current-measuring shunt and pack voltage interface.  Final specs include 9-18V power supply input, 500V maximum battery pack voltage, 2.5 uV shunt current resolution (17ma at 500A) and is Arduino IDE compatible for easy software development.  The enclosure will be waterproof and have an Amphenol connector similar to the GEVCU.  Since there is no display, for it to be useful you have to read the values with another device such as a GEVCU or any other device that can show the values on a user's device such as a smartphone or tablet.  Cost is expected around the same price as the JLD 404.

Next Session: EVTV Battery Management System "You've Got To Be Kidding" with Ed Clausen, Collin Kidder, Paulo Almeida, Celso Menaia and Jack Rickard

As anyone who has watched 5 minutes of EVTV video knows, Jack hates the "BMS" or Battery Management System.  These systems generally include a small circuit board attached to every battery that attempts to keep an even charge on each battery across the pack by reading voltage and bleeding off excess power in the higher-voltage cells as heat.  Unfortunately the failure mode for this kind of device usually results in burning down your car and whatever building it was parked in.  This is bad, because the device that's supposed to protect your battery pack actually destroys it.  Living without a BMS is tricky because you don't know the health of each cell, or at the least, each group of cells in the pack.  Jack solves this by carefully bottom-balancing each battery to within a thousandth of a volt and then the nature of LiFePO4 chemistry ensures that the batteries exactly track each other during the charging and discharging phase.  No inter-battery drift means no need to bleed off the high cells means no BMS is necessary.

So why is Jack now working on BMS systems with Paulo, Celso and Ed?  They're making Battery Monitoring Systems, not Battery Management Systems.  The PakTrakr system I had in my 914 was monitoring only, no active management.  This session will give us the first details on what Ed has been tinkering with in his shop.

The BMS will have 4 voltage inputs, giving the voltage of 4 segments of the pack, each being any size.  It will constantly sample the 4 segments and compare them, weighted properly, to see if things are getting out of sync.  The tricky thing is not acting like a parasitic load on some of the batteries and over time unbalance the pack, especially avoiding the ladder effect of multiple voltage measurements.

The heart of the system is a Sendyne SFP100 precision measurement chip, which has a very high reading resolution.  It can measure voltage, current, amperage and 4 points of temperature.  It has continuous calibration built-in to maintain the high accuracy.

The processor is the same ARM Cortex M3 as the GEVCU, and has EEPROM data storage, CAN bus, USB bus, I2C bus, digital interface to a high voltage multiplexer, 4 thermistors and an isolated power supply.  It will be housed like the GEVCU and have an Ampseal connector.  The voltage multiplexing lives on a separate circuit board with isolation on all inputs and I2C communication back to the main board.  The parasitic load is avoided by using a Bidirectional-blocking switch using two opto-isolated MOSFETs.  Each pair of MOSFETs is turned on and off in sequence.  A sampling capacitor takes about 600ms to charge up, then the ADC reads the voltage of that section of the pack.

I asked and Ed confirmed that it's theoretically possible to have one main board and multiple voltage boards, each with 4 voltage inputs due to the way he's multiplexing the voltage reading back to the Sendyne chip.  This way, people can get the amount of voltage sensing they want, from 4 points around the pack all the way down to each battery.

Ed's testing with the system involved charging with a Brusa charger and discharging through a grid-tied solar inverter.  He's seeing a lot of noise on the current sensing but that can be smoothed out in software.  Next steps include writing software for the board rather than the Sendyne test app, integration with the GEVCU, a slight redesign of the circuit board to account for a non-standard package size on the MOSFETs and creating an enclosure for the main board.

After the sessions we went back to the EVTV workshop.  The long Karmann Ghia project got a helping hand from some EVTV attendees and got it wired up and running.

The Smart car is charging and very close to being rolled off the rack and driven down the road!

And that wraps up the day!  Tomorrow we head over to the airport to get liquored up, play with high voltage and drive fast cars!